OhhO Product Video Prompts — Higgsfield.ai

AI-video generation prompts for a product demo / introduction film for every OhhO product, written for Higgsfield.ai — the platform these films are generated on. The prompts are built around what Higgsfield does best: multi-shot storyboard films sequenced from 15-second shots, photoreal model output, and an unlimited plan that makes long, many-shot films free to iterate on.

Source of truth for what each product does: ../products/. The look is deliberately photoreal live-action — NOT the website's current UI or a CGI/motion-graphics style. The site's UI will change; the films must not date with it. Screens in frame are always soft and out of focus.

The files

The 21 product prompt files: build · frame · bench · connect · bridge · link · ohho-os · serve · view · data · train · autonomy · mind · market · pilot · fleet · twin · care · comply · shield · proof

How Higgsfield works (and how these prompts use it)

Higgsfield is a multi-model AI video platform. The relevant capabilities:

Model picks (in order of preference for this campaign):

Use whichever of these your plan's unlimited selection covers — the prompts are model-agnostic on purpose.

The look: photoreal live-action (not CGI, not the site's UI)

Earlier versions of these prompts rendered a stylized 3-D "glass console" UI. That is retired. The films are now photorealistic live-action: real robots, real hands, real benches, cinema-camera grammar. Three rules keep them on-brand and future-proof:

1. No readable UI anywhere. Every monitor, tablet or terminal in frame is a soft out-of-focus glow. The website's UI will change; the films won't date.
2. Brand color lives in practical light. Cyan #00D4FF and violet #7C3AED appear only as physical light sources — LED strips, status rings, screen spill — never as graphics. Each film is led by one of the two accents.
3. It must read as filmed footage. The NEGATIVE list bans the CGI/render look, holograms, wireframes and motion graphics outright.

The STYLE block (identical in every prompt — do not edit)

This exact text leads every shot of every film. It is reproduced here so you can verify it never changes between products:

STYLE — OhhO (identical for every film, do not change):
Photorealistic live-action cinematography — a premium robotics brand film shot like an Apple product documentary. Real physical robots and real hardware in a modern robotics lab / light-industrial space: polished concrete floor, dark matte walls, aluminum-extrusion workbenches, neat cable runs, tool walls, server racks and large monitors. All light comes from practical sources inside the space: one cool white key light, cyan LED strip accents (hex #00D4FF) and deep violet ambient fill (hex #7C3AED), with thin atmospheric haze catching the beams. Cinema-camera look: full-frame sensor, 35mm and 50mm prime lenses, shallow depth of field, filmic highlight roll-off, subtle natural film grain. The camera moves only on a motorized slider, gimbal or jib — one slow deliberate move per shot, no shake. Materials must read as physically real: brushed aluminum, anodized black metal, rubber wheels, braided cables, glass screens with true reflections and fingerprint-level detail. Every monitor or tablet in frame shows only a soft, out-of-focus dark dashboard — never readable text. Mood: calm, confident, precise, quietly futuristic. Color grade: cool and high-contrast, near-black shadows with neon cyan and violet practicals, never oversaturated.
NEGATIVE: no CGI look, no 3-D render or motion-graphics aesthetic, no floating holograms, no wireframes, no cartoon or anime, no recognizable human faces (hands, silhouettes and out-of-focus figures are allowed), no readable on-screen or printed text, no company logos or watermarks, no warm sunny tones, no oversaturation, no cuts inside a shot, no shaky handheld camera, no heavy lens flares.

Each film then adds a one-line PRIMARY ACCENT paragraph (cyan-led or violet-led practicals) and its numbered SHOT blocks.

The shot grammar

Every product film is 6 shots × 15 s ≈ 90 s (the hero demo.md is 8 shots ≈ 2 min), and every film follows the same arc so the campaign cuts together:

SHOT 1  Establish     — the space, the hardware, the accent practicals, haze.
SHOT 2  First action  — the product's story begins with one concrete physical act.
SHOT 3  The mechanism — the signature capability, shown not told.
SHOT 4  Depth beat    — a second angle on the capability (scale, precision, contrast).
SHOT 5  The differentiator — the beat only this product can claim (recovery,
        swap, verification, publish…).
SHOT 6  Hero          — a slow orbit/push on the resolved scene; accent light
        breathes once; the UPPER THIRD is held as clean dark negative space
        for the title overlay. Ends on a hold (extender-friendly).

One camera move per shot, named explicitly (slider, dolly, gimbal arc, jib, tracking, orbit, push-in) — matching Higgsfield's camera presets.

Make the photoreal keyframe still first (recommended)

Generate one still with Higgsfield's image model and reuse it as the first frame of SHOT 1 across all films:

The full STYLE paragraph above, composed as a single hero still: a modern
robotics lab in near-darkness, polished concrete,
one workbench rimmed by a cyan LED strip, a mobile robot silhouetted in haze,
violet ambient fill deep in the background, cinema-lens shallow depth of field,
photoreal. The upper third of the frame is clean dark negative space. No text,
no faces. 16:9, photorealistic, filmic grain.

Save it as ohho-photoreal-keyframe.png and attach it as the image-to-video first frame. It replaces the old CGI "brand keyframe".

Higgsfield workflow (per film)

1. Open the product's .md file; copy the full storyboard block.
2. Higgsfield → Create → Video → storyboard / long-video mode → model Seedance 2.0 (or your unlimited model) → 16:9 · 1080p · 15 s per shot.
3. Attach ohho-photoreal-keyframe.png as SHOT 1's first frame; fix a seed where the model exposes one and reuse it across films.
4. Generate. On the unlimited plan, regenerate individual shots freely and keep the best take of each — takes are free, taste is the budget.
5. Chain: storyboard mode handles transitions; generating shot-by-shot, feed each shot's last frame as the next shot's first frame.
6. Stretch: use the AI Video Extender on any shot that needs more air, and loop the hero shot for ambient/background use.
7. Assemble in your editor: shared 2 s wordmark sting at the head, the product's text overlays (below) on the hero shot, shared CTA card, one shared LUT.

Post-production: the text overlays (add these in your editor)

Never let the model render text. Composite these exact strings over the hero shot's negative space using the brand fonts and hexes.

Brand fonts: Space Grotesk (wordmark + product name), JetBrains Mono (eyebrow labels, CTA, mono UI), Inter (any body line). Hexes: background #0A0E1A, cyan #00D4FF, violet #7C3AED, white #FFFFFF. Wordmark rule: render OhhO with the first and last O in cyan #00D4FF and the middle hh in white — exactly as on the site. Always end on The Open Robotics Platform.

Suggested overlay timing (90 s film): 0–2 s OhhO wordmark sting (shared) → footage clean until the hero shot (≈75 s) → product name + tagline fade in on the negative space → final 5 s adds the CTA and The Open Robotics Platform.

Sound design

Seedance 2.0 and Veo generate native audio — either keep the diegetic sound (servo whirs, fan hum, connector clicks suit the photoreal look well) or mute and lay one shared bed in the edit: a low warm pad, room tone, real mechanical foley on the action beats, one soft sub swell on the hero light-breathe. No music with lyrics. Same bed across all films.

Accent + category reference (which color leads each film)

Embedding the videos on the product pages

Each product detail page renders a glass-card "console" slot — that is where the film goes. In website/components/products/ProductDashboard.tsx (rendered by app/products/[slug]/page.tsx), the per-slug mockup can be replaced with the generated film:

<video
  className="w-full h-full object-cover"
  src={`/videos/products/${slug}.mp4`}
  poster={`/videos/products/${slug}.jpg`}   // the photoreal keyframe still
  autoPlay muted loop playsInline
/>

Two exports per product:

• Card loop — the hero shot alone (15 s, or extender-looped), muted H.264/WebM, for the autoplay slot above: public/videos/products/<slug>.mp4.
• Full film — the complete ≈90 s cut with overlays and sound, for the product page's expanded player / social: public/videos/products/<slug>-full.mp4.

The 2-minute demo film exports to public/videos/demo.mp4 (full) plus a looping hero-shot cut for the homepage background.

TL;DR workflow

1. Read this file. 2. Make the photoreal keyframe still. 3. Open a product file (e.g. serve.md). 4. Paste its storyboard into Higgsfield's storyboard / long-video mode — Seedance 2.0, 16:9, 1080p, 15 s per shot, keyframe as first frame. 5. Generate; on unlimited, re-roll shots until each take is right.
2. Extend / loop with the AI Video Extender where needed. 7. Add the shared sting + overlays + LUT in the edit. 8. Export the card loop and the full film to public/videos/products/. Repeat for all 21 + the demo — the locked STYLE block and shared keyframe keep them one campaign.